Research Gate

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/241727030

The Ice Cream Stories: A Study in Normal and Psychotic Narrations

Article in Discourse Processes · July 1986

DOI: 10.1080/01638538609544645

CITATIONS READS
15 207

2 authors, including:

Elaine Ostrach Chaika

Providence College
32 PUBLICATIONS 678 CITATIONS

SEE PROFILE

All content following this page was uploaded by Elaine Ostrach Chaika on 18 March 2014.

The user has requested enhancement of the downloaded file.

I---
DISCOURSE PROCESSES 9, 305-328(1986)

The Ice Cream Stories: A Study in

Normal and Psychotic Narrations
ELAINE CHAIKA
Providence College

PAUL ALEXANDER
Brown University

An adaptation of the Pear Stories study originally undertaken by Chafe and his collab-
orators ( 1980) was presented to 22 psychotic patients with discharge diagnoses of schizo-
phrenia and mania, and 25 normal controls. The psychotic and normal populations
showed definable differences in encoding strategies. Psychotic narratives evinced defects
in narration ranging from serious neologizing, disruptions in syntax, and narrative tech-
nique. Normals and psychotics showed some mutually exclusive dysfluencies. The strat-
egies of normals caused them to misperceive certain events. It could not be unam-
biguously determined whether the psychotics had actually misperceived or whether they
had only misencoded. This study does not support theories which claim that psychotics
have intact linguistic abilities. It does support theories claiming faulty filtering mecha-
nisms, vulnerability to distraction, and attentional deficits, all of which seem to be
referring to the same phenomena.

INTRODUCTION
..
At what point does discourse become crazy rather than creative? What makes
listeners judge one narration as being within the bounds of normalcy and another
beyond the pale? Is there an underlying mechanism or defect in linguistic-
cognitive processing which can explain the abnormality in psychotic speech? If
so, how can we uncover it? To answer these questions an adaptation of Chafe's
(1980) study of narration, The Pear Stories, was developed for a psychotic

Part of the research for this paper was undertaken with the help of a stipend for a summer seminar
from the National Endowment for the Humanities to the first author. Many thanks are due to
Professor Joseph Williams who first suggested that Chafe's work was pertinent to my interests.
Thanks are also due to Professor Richard Lambe who strengthened my conclusions with his statistical
magic on respective narrative lengths. Deborah Tannen's careful suggestions improved the entire
paper. Any errors remaining are my own. Finally, thanks must go to Professor Roger Desautels,
Director of the Providence College Audio-Visual Lab, who provided both personnel and expertise in
the filming and editing of The Ice Cream Story. An earlier version of this paper was presented at
the annual meeting of the Linguistic Society of America, December 28, 1982, in San Diego, Cali-
fornia.
Correspondence and requests for reprints should be addressed to Elaine Chaika, Professor of
Linguistics, Providence College, River Ave., Providence, RI 02918.

305
306 CHAIKA AND ALEXANDER

population. Testing these results against normal controls yielded fresh insights
into the problem of "crazy talk." It will be shown that there are very real
differences between the two populations in narrating a simple story, and that
these differences show true linguistic dysfunction.

THE PROBLEM
Frequently, manics and schizophrenics talk crazy only during psychotic bouts,
but what this crazy talk means and why it occurs has had many interpretations.
With few exceptions (Andreasen & Powers, 1974; Chaika, 1977; Rochester &
Martin, 1979), the problem has been seen as one of schizophrenia rather than
mania, but, as this paper shows, to some degree it may be one of mania as well.
Hence, the study reported on here speaks of psychotic rather than schizophrenic
speech. Psychiatrists have seen schizophrenic speech in various lights: excep-
tionally veiled language obliquely referring to homosexual or other tabooed
desires (e.g., Laffal, 1965; Searles, 1967). Yet others have seen it as unusual
poetry (e.g., Forrest, 1965, 1976).
Psychologists have seen schizophrenic speech as a disruption in ability to
make correct word associations (e.g., Chapman, Chapman, & Daut, 1976), an
inability to retain fixed constructs (Bannister, 1962) or to "pigeonhole" cor-
rectly (Broadbent, 1971; Schwartz, 1982) or to filter stimuli competently (Neale
& Oltmanns, 1980; Reed, 1970). There are even those who see no structural
abnormality in schizophrenic speech (e.g., Brown, 1973; Cohen, 1978; From-
kin, 1975; Kertesz, 1982) claiming that no matter how odd such speech may
seem, the peculiarities are caused by thought disorder or some other factor
extraneous to competence in langvage.
Some of the above theories fail because the deficit posited has little if any-
thing to do with normal language functioning, hence can not be assumed to
underly dysfunctioning. Theories of response biasing or pigeonholing fall into
this category. Others fail because they belie the available data or do not account
for them.
The reason that we must ensure that our explanations take into account the
skills needed for normal speech production is that speech readily labeled
"schizophrenic" is evident in ordinary interactions. It appears that there may be
a disruption of normal linguistic abilities. Therefore, to explain such disruption,
we must test for those skills necessary for daily speech. In support of this
contention, it has often been reported that there is high interjudge reliability in
discriminating between normal and schizophrenic speech, and that lay judges can
discriminate between such speech and normal as well as psychiatrists can (Ker-
tesz, 1982; Maher, McKeon, & McLaughlin, 1966; Rochester, Martin, &
Thurston, 1977). We found that this extends to these narratives as well. In this
experiment, two raters, judging the narratives from a tape while reading its
-
PSYCHOTIC NARRATION 307

transcript, considered three narratives from normals to be psychotic. As shown

below, these did, indeed, pattern with the psychotic narratives in at least one
way. Two other normals were rated as psychotic by one judge, and, again, there
was a definable reason. Only one psychotic was deemed normal by both raters,
and he was taped near the end of his hospital stay.
Because we can correlate judgments of abnormality with specific features of
narratives, studies such as this one, which elicit narration, should lead us to an
understanding of what features of a narrative cause it to be heard as abnormal in a
given culture.
Several excellent reviews by psychologists and psychiatrists exist of the copi-
ous explanations for schizophrenic speech (e.g., Maher, 1972; Reed, 1970;
Rochester & Martin, 1979; Schwartz, 1978). Chaika (1974, 1977, 1982a,
1982b) has reviewed the literature on schizophrenic speech from the viewpoint of
a linguist on the basis of what is known about normal and artistic language use.
Chaika (1982a) shows that the failings in schizophrenic speech are largely,
although not entirely, failures at the level of discourse. Whereas some disclaim
such failures as being merely pragmatic (Fromkin, 1975; Laffal, 1982), hence
not a true language deficit, others, such as Martin (1982), make it plain that
pragmatics must be considered part of language. Indeed, it can be-and has
been-shown both by philosophers such as Austin, Grice, and Searle, as well as
sociolinguists like Philips (1976) and Sacks (1964-1972), that meaning resides
in pragmatics on the part of both the speaker and the hearer.
Because language is constructed to give meaning, and because meaning can-
not be interpreted outside of context, pragmatics, which shows how things mean
in context, cannot be ignored as irrelevant to speech dysfunction (Martin, 1982).
It is the position of this paper that pragmatics is an integral part of linguistic
competence.
The validity of any explanation for the apparent deviations in schizophrenic
speech ultimately rests upon the relationship of what has been said as opposed to
what it is that is being encoded. Indeed, much of the discussion-even con-
flict-over what schizophrenic speech means has centered on that very issue.
When we can't fit schizophrenic utterances within a given context, is it because
they are responses to hallucinations? Is it because the schizophrenic is speaking
about his or her peculiar inner world, trying to explain the schizophrenic experi-
ence (Forrest, 1976)? Is it an attempt to avoid the therapeutic situation (Laffal,
1965)? Or, is it because the schizophrenic has diminished competence in lin-
guistic production at least during the time he or she is speaking bizarrely?
This paper attempts to show both that (a) psychotics have a genuine disruption
in their ability to produce a competent narrative, and that (b) this disruption is
caused by an apparent disability in organizing language at any or all of its levels,
from word finding to sentence constructing to narrative constructing. This in-
ability seems caused in tum partially by an inability to suppress personal memo-
-
308 CHAIKA AND ALEXANDER

ries or words and phrases, including cliches inappropriate to the task at hand
(Chaika, 1982a). Many theories advanced for the oddities of schizophrenic
speech have discussed its strange associational character, including this mix of
memories with other verbal output. terms like "filtering defects," "faulty
pigeonholing," "attentional deficits," and "weakening of constructs" have all
been used both as explanation of the cause of such language and as a description
of it. All of these terms seem to be referring to the same phenomena. This study
indicates that schizophrenic narration is marred by intrusions from personal
memory, such that it seems to be suffering from ''faulty filtering'' mechanisms.
It should be stressed, however, that other terms might be-and have been-used
to label the same phenomena.

THE PEAR STORIES

Because of these issues, the methods described in Chafe (1980) are an excep-
tionally good vehicle for testing assumptions about schizophrenic and manic
speech, henceforth labeled psychotic speech. Chafe and his coworkers had peo-
ple view a short movie with no speech. Then, a researcher asked them to narrate
what they had seen. The beauty of such a procedure is that the narration can be
matched to what the participant is trying to encode. This task also demands
narratives requiring the encoding of ongoing events, which in tum demands
temporal sequencing and shifting references, thus simulating everyday narrative
situations.
Assuredly, this is not the only experimental technique which, in essence, has
correlated speech production with the object of encoding. Rochester and Mar-
tin's (1979) study, for instance, had schizophrenics describe cartoons and retell
an anecdote. Cohen (1978) had patients describe color chips, and many others
have collected data by correlating answers to questions asked. All of these
studies have yielded valuable results, but none directly comparable to those from
Chafe's procedure.
Explaining the points of cartoons, retelling anecdotes, and describing unusual
colors are not everyday activities and would seem to present a greater cognitive
load than a mere retelling of an ordinary sequence of events. The problem with
questioning without strongly constraining the responses is that, too often, psy-
chotics wander off the point. As a result, it is sometimes difficult to correlate
utterances with intended meaning. As Maher and others ( 1966) and Maher
(1972) noted, the more unconstrained the speech activity, the more disorganized
schizophrenic speech becomes.
One kind of disorganization, indeed one considered peculiarly schizophrenic
(Lecours & Vaniers-Clement, 1976; Werner, Lewis-Matichek, Evans, & Lito-
witz, 1975) is glossomania, typically a string of phrases or clauses, related
primarily because individual words either share syntactic, semantic, or pho-
 PSYCHOTIC NARRATION 309

nological features with each other. An example (Chaika, 1974) of semantic

glossomania is:

(a) My mother's name is Bill-and coo? St. Valentine's Day is the start of the
breedin' season of the birds. I like birds. Especially parakeets. They work
hard ...

Here, the relationship between the phrases is determined by the chance associa-
tion of Bill with the expression bill and coo, which is an old idiom for "love." In
turn, this prompts mention of the holiday which celebrates love, but the original
literal meaning of bill and coo, something which birds do, prompts the mention
of birds.
An example (Cohen, 1978) of both semantic and phonological glossomania
is, in response to being asked what color a chip is:

(b) Looks like clay. Sounds like gray. Take you for a roll in in the hay. Hay
day. Mayday. Help. I need help.

Here, the chance similarity of the syntactic constructions for verbs of sensing,
seems to prompt sounds like on the model of the previous looks like. The gray
seems to be chosen because it rhymes with clay. The next sentence is evoked
because of the continuing rhyme of clay, gray, and hay. However, hay day
seems to be uttered both because of the rhyme and of the meaning of "fun" in
both a roll in the hay and hay day. Mayday continues the rhyme, but help seems
invoked by the meaning of mayday.
Disorganized discourse such as that in (a) and (b) above frequently gives us
valuable information and certainly must be accounted for in any explanation of
schizophrenic speech. However, it is also necessary to be able to correlate
utterances with rather more precision to what it is ostensibly that schizophrenics
(and other psychotics) are trying to encode. Because of the strong constraints put
on responses in the task reported on here, glossomanic chaining, for instance, in
its above-shown "classic" forms did not occur, although a variant of it did.
What is perhaps most surprising is that patients did utter both gibberish and
agrammatisms despite the constraints on the task, and despite claims of re-
searchers that such aphasia-like symptoms are rare in schizophrenia (e.g., Co-
hen, 1978; Lanin-Kettering & Harrow, in press; Schwartz, 1982).

THE METHOD: THE ICE CREAM STORIES

Given the short attention span which typically accompanies psychosis, the prin-
cipal investigator determined that the Chafe movie, although very short (about 6
min) was too long. Also, because it dealt with stealing and a bicycle accident, it
310 CHAIKA AND ALEXANDER

was felt to be potentially upsetting to patients. Then, too, Chafe did not include
dialogue in his movie as he wanted to use it with speakers of many languages.
Because one of the issues which has been raised with schizophrenics and manics.
is whether or not they understand utterances as normals do (Chaika, 1977;
Laffal, 1965), it was decided to include dialogue. Finally, fearing that the para-
phernalia of movie projection might prove too distracting, the principal investi-
gator made a 124-s videotape. This was played through a 12-in. JVC monitor
which looked exactly like an ordinary television set.
Because Butler Hospital, the location of the experiment, had a lounge with
couches available on the Intensive Treatment Unit, the entire experience simulat-
ed an experience likely to be familiar to all patients: that of watching a television
show and then describing what they had seen.
There was another departure from the Chafe experiment. Because they were
showing a movie, they had several people view at once. Then, the viewers were
called one by one to recount what they had seen. Neither the 12-in. monitor nor
the condition of the patients allowed such multiple viewing of The Ice Cream
Stories. Therefore, each viewed the story with only the principal investigator
and, in some instances, a mental health worker present. Then, each patient was
asked to tell in his or her own words what he or she had just seen. Thus nobody
had time for a narrative to cook, so to speak, accounting for at least one interest-
ing parallel between narration and eye movement, as noted below.

THE STORY OF THE ICE CREAM STORY

Because actively psychotic patients frequently have a short attention span, The
Ice Cream Story could not be as complex as the Pear Story. Still, some extra-
neous material was included because one of the characteristics of the speech of
schizophrenics frequently mentioned in the psychiatric literature in their veering
from the topic at hand. This is termed derailment. It is considered so prevalent
that it is one criterion by which schizophrenics are diagnosed. Given hypotheses
about the nature of schizophrenic malfunctioning in attention and filtering, it was
expected that the extraneous material would cause derailment. Actually, when
derailment occurred, it was from the essential plot and appeared to be caused by
intrusions from memory, as is shown below.
The story itself not only is simple, but relates an incident familiar to most
Americans. The first scene pans a shopping center, closing in on the figure of a
little girl looking through the window of a Baskin Robbins store. The next scene
shows a woman setting a table, and the same girl walking into the room asking,
"Mommy, can I have some ice cream?" whereupon the mother leans down, puts
an arm around her and says gently, "No, honey, it's too close to suppertime."
Then a man is seen walking into the house. The child walks up to him, touching
her body to his. He says, ''Hello, Stefanie. '' Then she asks, ''Daddy, can I have
 PSYCHOTIC NARRATION 311

ice cream?" The father looks into the camera with a grin, and his hand moves
towards his pants pocket. The next scene shows the child walking towards the
Baskin Robbins store, entering, leaning against the counter as she waits fidget-
ing. Then a clerk comes into view, asking it he can help her. She responds
inaudibly, but the man repeats clearly, ''Double grape ice.'' The child plays with
coins, still leaning on the counter. The man returns with a very large double-
decker cone. The girl gives him the money which he looks at, then rings up on
the register. A bell chimes on the register. The man gives her change, and says,
"Thank you. Come again." The girl turns towards the camera with a triumphant
smile, pushes the door and goes out. A sound of "Oh, wow" comes from
outside the door. The film ends there.
Despite its simplicity, the Ice Cream Stories tested for many language skills
and attentional and logical phenomena. For instance, the viewer had to leap one
important gap. The father is not actually seen giving the child money, nor does
he answer her in words. When one sees her walking in to buy the ice cream, one
might surmise that the father must have given some money to her. Scenes such as
one showing the mother setting the table were included to see if they would cause
the narrator to be deflected from the major progression of action, perhaps starting
off on something relative to mealtimes or mothers or family incidents-as one
might expect given the prevalence of glossomania in schizophrenic speech
(Chaika, 1982a; Lecours & Vaniers-Clement, 1976). As indicated above, how-
ever, deflection caused by these side actions did not happen.
The 124-s story proved to be well within the attention span of all participants.
The opening shot, panning a child wearing a plaid skirt and vest with long-
sleeved jersey peering into the window of a Baskin Robbins, took 20 s. Later,
when the child enters the ice cream store, it takes 23 s for her to be waited on. In
terms of effects on the narrations, these seemed to be the only significant time
spans.

CHARACTERISTICS OF PARTICIPANTS IN THIS STUDY

A major problem compromising research into the speech of schizophrenics is the
sporadic and cyclic nature of deviations in their speech production. Many studies
have selected patients solely on the basis of diagnosis without ascertaining
whether or not they were part of the subset of patients who evinced structurally
deviant speech at any time. This has contributed to the dissension about whether
or not schizophrenics suffer from a lapse in linguistic competence (Chaika,
1981). If one happens to be dealing with those who speak structurally normally,
one can say as does Bertram Cohen (1978, p. 1), for instance, "hard instances of
agrammatism" don't occur in schizophrenia, or find, as did DiSimone, Darley,
and Aronson (1977), that, on an aphasia test battery, schizophrenics do not
respond as aphasics do. Then, too, the causes of schizophrenic, aphasia-like
312 CHAIKA AND ALEXANDER

utterances may not respond to such testing because schizophrenic surface anoma-
lies may be caused by different deep-structure disrupticns. In any event, hard
instances of agrammatism and of disruption in speech to the point of gibberish
were not hard to find in this experimental proce<lure.
Because this study is concerned with structurally strange speech, not neces-
sarily strange content, mental health workers on the Intensive Treatment Unit
were briefed to note patients who evinced some of the features associated with
schizophrenic speech: glossomania, neologizing, gibberish, opposite speech,
inappropriate rhyming or punning, word salads, perseverations, or faulty cohe-
sion (Chaika, 1974, 1982a). It was not expected that any one would produce all
or even most of these, and, in fact, nobody did. As with other disrupted speech
most of each narrative was decodable, albeit not necessarily by the usual strat-
egies for comprehension (Chaika, 1977).
The preselected patients were then invited to participate in the study. Of the
original 24, 2 were dropped from the study because it was discovered that they
probably had drug-induced psychoses. This selection process yielded those with
discharge diagnoses of mania as well as schizophrenia. Diagnoses were made by
the attending psychiatrist, Paul Alexander (coauthor, this volume), on the basis
of DSM II and III.
All patients were being treated with antipsychotic medication as well as
antiparkinsonian medication designed to mitigate the side effects of the former.
All also were receiving lithium (Alexander, VanKammer, & Bunney, 1979).
The effect of these medications is to lessen the effects of psychosis, so that
speech is made more normal. This makes even more important the very real
differences found between the normal and psychotic narratives.
The average stay at the hospital during the time of this study was less than 2
weeks. No participant had been institutionalized or heavily medicated for long
periods. Hence our results could not be traced to social or cognitive deficits on
those grounds.
Twenty-five normal volunteers matched the psychotic population as closely as
possible in age, occupation, and social class. Both groups consisted of blue-
collar workers and college students. As is usual between psychotic and normal
groups, the normals were somewhat higher in achievement. This effect was
mitigated by including normal college students with working-class parents and
psychotics with college-educated parents. Bernstein's theory of differential nar-
rative adequacy rests upon a theory of socialization so that, if it were valid,
parental status should be as important as earned achievement, especially for
nonachieving children of educated parents. One could argue that those who rise
from the working class do so because they somehow learn on their own what
Bernstein calls the elaborated code (Bernstein, 1971). However, it is difficult to
claim the reverse, the scions of the middle class sink, as it were, because they,
despite their socialization, only learned a restricted code, for Bernstein's claim is
 PSYCHOTIC NARRATION 313

that the latter is different in kind from the former, not just in degree (Chaika,
1982c, for further arguments against Bernstein's theories). In any event, the data
presented here correlated with diagnosis of psychosis versus normality, not with
social class.
The three normals judged psychotic by both raters -were all college educated.
Their narratives showed definite correlates with the psychotic ones. Rochester
and Martin ( 1979), who claim that, in their experiment, schizophrenics use more
exophora to refer to items or people in the cartoons shown, liken this to Bern-
stein's theory of restricted code amongst the working classes. Such a correlation
was not found in this study. The difference in results may well have been caused
by the differences in the experimental protocols (Chaika, 1983; Chaika, Lambe,
& Alexander, submitted). The Ice Cream Story task has subjects narrate after the
scenarios are no longer in view so that exophoric reference to them is not as
feasible. Rochester and Martin's subjects spoke while the pictures were in view.

THE QUESTION OF COOPERATION

It is especially important to establish that the psychotic population was trying to
fulfill the experimental task. It may be argued, and has often been, that such
patients produce deviant discourse because they wish to, either because they
want to confound the investigator, or because they are especially creative
(Chaika, 1974, 1977, 1981, 1982a). Alternatively, one might argue that the
psychotic participants in this study failed because they did not understand what
was expected of them. If, indeed, they were not cooperating or if they did not
understand the task, then our results would be meaningless, because these rest
wholly on the correlation of the narrative to the videostory.
Searle ( 1983) has shown that Intentionality is always an integral part both of
speech production and of meaning itself. He (p. 3) defines Intentionality as
directedness, and shows that meaning is comprised of " ... Intentional content
that goes with the form of externalization" (Searle, pp. 28-29). Thus normal
comprehension demands that we derive Intentionality as well as truth value
(Chaika, 1982c, p. 71; Goody, I 978). If speech is so deviant that we can not do
so, then we fault the speaker. If the speaker does not encode so that his or her
Intentionality (or truth value) can be derived, and, if further, the speaker can not
explain when asked, we fault the speaker, not the hearer. Hence, we consider the
opaque or deviant speech of psychotics as evidence of a person's being "out of
his/her mind." Similarly, as Searle shows with visual perceptions, when people
see things that are not objectively there, we say " ... it is the visual experience
and not the world which is at fault" (p. 43).
Searle (1983, p. 147) introduces another component to meaning, one es-
pecially valuable in this discussion: the Background. He defines this as '' ... ca-
pacities and social practises." Hence, Background includes " ... skills, prein-
314 CHAIKA AND ALEXANDER

tentional assumptions and presuppositions, practises and habits" (Searle, 1983,

p. 154). Under this view, a breakdown in surface performance is a failure of
''. . . preintentional capacities that underlie the intentional states in question''
(Searle, 1983, p. 155). To treat that which arises from faulty Background as if it
were " ... a sort of Intentionality, it immediately becomes problematic" (Sea-
rle, 1983, p. 159). What this means here is that we assume that narratives which
are deviant either in what they report or the way that they report arise from
impaired skills in narration, not from a separate language or an attempt to hide
taboo desires or an attempt to convey what it means to be schizophrenic or the
like. This is the simplest explanation that fits the facts and is the one adopted in
the explanations below.
It is still necessary to show that the participants in this study understood the
task, were cooperating in it, and intended to fulfill it. There are five reasons for
believing this.
First, even those whose narratives were not accurate took as their point of
departure the sequence shown. Moreover, when they digressed from this, they
related stories similar to that on the screen, and kept returning the events
(boldface) actually shown them, as in (1):

(I) What do you want me to say? I saw my brother Gene. He says he said I buy the
things I wanted. I saw a little girl who wanted ice cream. Today you have to pay
for it. Today she paid for it.

Second, where someone hallucinated or misperceived action, he or she indi-

cated an effort to integrate it into the story. In (2), for instance, when the patient
says, "I don't know what ... that was about," he indicates that he can not fit
what he perceives into the story:

(2) I saw a little girl who was moving a counter for some reason and I don't know what
the heck that was about. She was pressing against it okay. In the beginning I saw
a white car with a red vinyl top and then this little girl was looking in the store was
looking in the trash can or something and then she turned around and she went on she
talked to her mother and her father and neither one of them was listening to her . . .

Here, note also the "okay" as indicating that although he didn't "know what the
heck that was about" he was going forth with the narrative anyway, and he
proceeded to recount the tale on the tape.
Third, several made overt comments about their ability to speak or to re-
member something on the tape. For instance, one apologized:

(3) ... she just cunna's cunna get anything home so she's hafta go out on her and get it.
okay. I'm sorry, I'm sorry. I'm sorry about my speech. I stutter a lot though.
That's about it.
 PSYCHOTIC NARRATION 315

Indicating a ''memory lapse'' also shows an effort to recount what was shown:

(4) . . . and I noticed a little girl looking into the window and I guess he walked back
into the store and then a [kif] thing switched where the girl was at home and I dunno
I asked her mother for something and she had a kni- got a little memory lapse there.
Then it switched again and her father came in . . .

That patients were attempting to narrate the videostory was also indicated by
self-corrections, and by expressing happiness that the girl got her ice cream:

(5) so then she went and she went to the candy store by herself or ice cream shop and
bought a double-decker ice cream cone. That actually brought me happiness to
see that little girl with a mind of her own. Okay.

Even the most deviant narratives signalled endings formally, marking the
narrative as an entity. The "That's about it" in (3) above is one example, as is
the "Okay" in (5).

RESULTS: VISUAL SCANNING

The parallel between visual scanning and narration made by Chafe ( 1980) was
borne out in this study. His claim is that the narrative itself progresses in a
fashion similar to the way the eye searches a scene. Chaika (l 982a) correlates
schizophrenic dysfunction in visual tracking, "spiky-type" eye movements
(Holzman, 1978), with dysfunctions in their free speech. Chaika ( 1982a) showed
that such spiky-type movement is analogous to schizophrenic utterances inap-
propriate to the context, but explicable by assuming "spiky-type" movement
along associative pathways, as in (1), here repeated:

(l)· What do you want me to say? I saw my brother Gene. He says he said I buy the
things that that I wanted. I saw a little girl who wanted ice cream. Today you have to
pay for it but today she paid for it . . .

Note that the first question is entirely appropriate as a narrative opener. Howev-
er, the story line is intertwined with her memories and desires so that she ''sees''
both her brother Gene and the videostory. As with spiky-type eyetracking, she
follows first her inner story about Gene, then picks up the videostory, then jerks
back to Gene, then back to the girl.
In (2) we see a visual analog to inappropriate schizophrenic punning (Chaika,
1974): "I saw a little girl who was moving a counter for some reason and I don't
know what the heck that was about." The girl's stance, leaning forward against
the ice cream case as she is waiting to be served, is similar to that when one is
316 CHAIKA AND ALEXANDER

moving a heavy object, and the narrator gave the wrong interpretation to this
stance.
There was also a verbal parallel to the visual process upon first seeing a scene
and not having a frame to put it in. Not knowing what is going to be relevant, the
person tries to note everything that is going on until he/she figures out a frame
for the unfolding scenario. Once this frame is constructed, only matters relevant
to the story line get mentioned. Because all were recounting the story soon after
seeing it, it is as if their initial narration was wholly unedited, so that they
verbally recounted the visual scanning upon first seeing a scene. Many started by
describing the cars and the people in the opening scene. Both populations fre-
quently did this. Note the similarity of the normal opener in (6) to the psychotic
one in (7):

(6) First I saw a parking lot with a lot of cars and I noticed an ice cream shop I think it
was a Baskin Robbins store. A woman walked by and another gentleman came from
the opposite direction and he walked past the screen and then I noticed a little girl
standing outside looking into the ice cream shop . .
(7) Okay. There's a lady who was walking toward the car and I forget it she was wa-
she walked by the car is what it was and they went past the car and a man walked by
a store a Baskin and Robbins sign it was the scene before so wa let's see then one
once they went past the man zooming in they they zoomed in on a girl . . .

Once they zeroed in on the girl staring in the window, however, they related only
those points of action that furthered the plot, typically that the child went home,
asked her mother for ice cream, the mother refused her because it was too near
suppertime, the father came home, the child asked him for ice cream, the child
went back to the ice cream shop and ordered ice cream which she received.
This "zeroing-in" tactic, also seen in Chafe (1980), is easily seen in the
degree of detail in description of characters. At the outset, many described the
clothing of the first characters seen, such as noting that "a man with a three-
piece suit minus the jacket" walked by, and/or a woman with a shopping bag
also walked by. However, such matters were never again mentioned again once
the narrator got his or her bearings. Not one person mentioned the clothing of
either parent, although each was on film far longer than the casual passersby at
the outset. Similarly, the opening parking-lot scenario was carefully described,
but the kitchen, which was important to the plot, received only one mention and
that, by a psychotic, commented on color: "There were cur- blue curtains. It was
kinda brown the room they're in." This scanning was not the only initial tactic.
Some immediately focused on the girl. Again, the two populations made sub-
stantially similar openings. For instance, compare the normal

(8) It began with a girl staring through a window at a Baskin Robbins store . . .
,---

' to the psychotic

PSYCHOTIC NARRATION 317

(9) I seen a little girl looking in the window and she want some ice cream . . .

However, after the openers illustrated by (8)-(9), differences between the two
populations became evident. Once normals got their bearings, so to speak, they
usually followed a strict temporal ordering in narration. They gave the impres-
sion of play-by-play description. First this happened, then that, and that, and so
on to the conclusion.
In contrast, psychotics showed no such orderly progression. Only 7 out of 22
followed a temporal order. The degree of misordering possible with psychotics is
seen in sequences like "She ... leaves the ice cream and eats it," and "She ate
the ice cream and brought it home."
One of the three normals deemed psychotic in both blind ratings was an
exception to orderly temporal scanning, as seen in (IO):

(10) A young girl getting ice cream at a ice cream parlor. Let's see what are the-and
there was a scene with her and her parents. She asked her father if he would give
her some money to get the ice cream and before that she was hanging around
outside the ice cream parlor. Okay let's see. How about she had a yellow shirt on
whatever. A sort of jumpsuit.

Another apparent exception to temporal ordering in a normal seems to have

been a slip of the tongue. One girl said, "He [her father] stuck his hand in his
pocket and then the film ended," but then proceeded to describe the girl going
back to the store and buying the ice cream. Apparently, what she was encoding
was that the scene in which the father is seen putting his hand toward his pants
pockets is abruptly cut off. She did not mean the narrative was through. She was
rated normal by both judges and her narrative conformed to the normal ones in
every other way.
Some psychotics showed an inability to "track" even after they had gotten
their bearings. For instance, they flitted from scene to scene, leaving out impor-
tant sequences, such as the fact that the girl first asked her mother for ice cream,
as in

(11) A little girl, she's uh she's on her own. She's so weh she gets her [ous oh)I after she
ask her own father if she can go out for ice ice cream and he says eh answers her shi
dunno and get ice cream for herself and es pass by sh wu and so it all happened that
they're all happy ...

1This and other "nonword" sound sequences were gibberish interspersed amongst recognizable

lexical items. This occurs in some patients, but not others; however, it is a well-known symptom of
psychotic speech.
318 CHAil(A AND ALEXANDER

or, as in (12), has her seeing ice cream before looking in the window:

(12) ... Well is this little girl that was looking in a-I suppose she was coming home
from school or something and saw an ice dish looked in the window and es saw
ice cream and it was an ice cream store and she would like to have some . . .

OTHER DIFFERENCES IN NARRATIVES

Psychotic deficits in temporal ordering seem related to a more general disor-
ganization reflected also in the gibberish in (11) above. Such disorganization
seems to cause other problems in the psychotic narratives, as noted below.
Although there was no significant difference in the length of narratives be-
tween the normal and psychotic populations overall, there was a difference in
what became encoded. The longest narratives are cases in point. Although each
of the following comes from narratives of roughly the !>ame length, they differ,
as one would expect from the above, not only in the progression of the story, but
in the amount of detail included. A comparison of the boldfaced passages in the
normal (13) as opposed to the psychotic (14) illustrates the poverty of detail in
the latter. Here the comparison is between the portion of each narrative concern-
ing the father:

(13) .. .she went up to her father who had just walked in the door and asked him if
she could have some ice cream which I guess is what she asked her mother and we
didn't hear her father's answer but then we returned to Baskin Robbins and
she walked into the door and ordered some kind of ice it looked like raspberry
and um the man . . . she waited for the ice cream cone. At this time her shirt
appeared yellow and the man gave her the ice cream cone; she paid for it and
left.

The psychotic narrative (14) lacks detail about what was shown. Rather, cliches
and interpolations lead it astray.

(14) ... He says well what the heck give it to her [nooee] sh- she's a little daughter so
he gets her the coins and she goes up ice cream stand and stands in line and
waits and gets a giant sized cone and she is so happy with her ice cream. A
simple pleasure, but that's what kids are like these days always have but th- it
means that [shinchuer] her parents that she's so proud of. She goes out leaves the
ice cream and eats it and on the way and we don't know what happens [sma] the
fact you can interpolate and say that she ate the ice cream and brought it home and
said thank you daddy thank you daddy or thank you mummy but she still is her
destination is not known in a few minutes . . .

Short narratives from each population show a similar disparity as can be seen
from the boldface in the normal (15):
 PSYCHOTIC NARRATION 319

(15) I saw a little girl looking into an ice cream store and she went home and asked
her mother if she could have some ice cream and her mother said no because it
was too close to supper and then she asked her father and her father gave her
the money and she went back to the ice cream store and bought some ice
cream . ..

Compare this with (16) from a psychotic:

(16) A little girl wanted things and her mother said no and her father came home
I·. and she asked for some ice cream and then she went back to the store and then
f. she ordered some ice cream and the man said thank you.
'

Not only did the speaker of (16) neglect some crucial elements, such as the
mother's reason for denying the request, and the father's response, but her
terminology is imprecise. The child does not want things, but ice cream, Similar-
ly, there is a strange gap between the girl's ordering the ice cream and the man's
saying "Thank you." In between, she paid for the ice cream and received it.
This is accurately encoded by the normal ·•She bought it,'' despite its brevity.

•
t
LEXICAL CHOICES
I'.
Psychotic narratives contained many emotionally laden words. The girl was said
to be "craving" ice cream, or was "put down" by her parents. In contrast,
normal language was usually colorless. There were only two exceptions to the
blandness of normal word choice. One normal said, "I saw her get rejected by
one [parent]. I saw one give in." Apparently, this caused one rater to judge this
normal narrative as psychotic. Another normal, judged psychotic by both raters,
said the girl "defied" her parents, also a strong word choice for a normal. It
... must be emphasized that such judgments are probably culture bound, as Tannen
(1980) reports that normal Greeks use such emotional words, although Ameri-
cans do not do so in this situation.
Occasionally, the unusual word choices of psychotics had an almost literary
or poetic sound to them as in (14)'s "He gets her the coins" or another psychot-
ic's "The cash register man handled the financial matters." However, these
felicities were rare, whereas actual errors in word finding were not, as shown
below. Rather than being evidence of exceptional creativity, as posited by For-
rest (1976) and Lecours and Vaniers-Clement (1976), these unusual encodings
may just have been accidental, a result of a general difficulty in getting the
correct word for the situation.

THE MORAL OF THE STORY

Not only did psychotics use somewhat more colorful language than normals, but
they were far more likely to comment on the point of the narrative. Some started
 .
320 CHAIKA AND ALEXANDER

with a psychological insight, "I seen a kid use mother versus father psychology
to get her ice cream and that's no just that she used psychology against her
mother didn't want her to have it." Another said, indignantly, "She was sittin'
there waiting for somebody to get musta waited two three minutes for get waited
on. A place like that should have it all the time soon as she comes in the door." It
is interesting to note that normals frequently commented on the fact that the child
played one parent against the other, even saying, "I always used to do that,
too." But normals made these comments only after the recordings were done,
thus showing that they did not perceive that such overt comments belong in the
narrative itself. ~
-
♦.

It is tempting at this juncture to hypothesize that psychotics are more con-

cerned with the abstract, or that, like Tannen's (1980) Greek subjects, they feel
impelled to give a moral. Because, in this task, Americans do not feel that they
should make a moral point out of what they see (Tannen, 1980), it is reasonable ..
to hypothesize that the psychotic moralizing we found is simply evidence of
lowered constraints on narration, a conclusion made more tenable by the fact that
other aspects of psychotic narration show such lowering. The veering off the
topic seen in (14) and the lack of precision in (16) are both evidence of lowered
constraints in the narrative situation.
Other aspects of psychotic narration support the conclusion of lowered con-
straints. Psychotics frequently seem to have difficulty suppressing material
which detracts from their utterances in a given narrative situation. That is,
associations not germane to the task at hand become encoded in their narratives.
In its least disruptive form these intrusions into narratives appear in asides like
"all about ice cream and I really coulda went for a cone," and saying that the
girl got a cone of grape ice, ''my favorite flavor.'' More disruptive intrusions are
seen in:

( 17) Okay. I was watching a film of a girl and um s bring back memories of things that
happened to people around me that affected me during the time when I was living in
that area and she just went to the store for a candy bar and by the time ooh of course
her brother who was supposed to be watching wasn't paying much attention he was
blamed for and I didn't think that was fair the way the way they did that either so
that's why I'm just asking yah could we just get together and try to work it out all
together for one big party or something ezz it hey if it we'd all in which is in not
they've been here so why you just now discovering it. You know they they've been
men will try to use you every time for everything he wants so ain't no need and you
trying to get upset for it. That's all. That's all.

Unlike the narration in (1) which at intervals cycles back to the business of the
film, this shows a progressive chaining. Like glossomania involving single
words or short phrases, here one association leads to another. In (17), individual
 PSYCHOTIC NARRATION 321

words do not seem to be the motivators for the associative chaining. It is as if the
next highest level of encoding, the sentence, is the governor. It might be that the
reference to the girl who went to the store for a candy bar was an erroneous
reference to the videostory. Even so, the rest shows no such returns to the
original. Note, however, that the deviation in this narrative, as in glossomanic
chains, arises from the lack of subordination to a topic (Chaika, 1982a).
Normals perceived the narrative task to be separate from their personal remi-
niscences. None of the normal narrations, even those judged psychotic by both
raters, 2 produced intrusions of glossomanic narratives.

NARRATIVE GLITCHES
Both normals and psychotics produced errors which interrupted flow in narra-
tion. They are here called glitches as they are analogous to glitching in vid-
eofilming terminology. Frornkins's (1975) assertion that schizophrenic error is
not different from normal error was not borne out, as there were three categories
of glitching produced only by psychotics, and one produced only by normals.
There was yet another error which only one normal but several psychotics made,
but this entire normal narrative was highly deviant. And, although both groups
made false starts, psychotics produced one type that normals didn't. 3
Both populations started a word, broke off, and then restarted, as in

(18) start/restart (a)f-f.for:,

(b) she we- went,"
(c) the way the way they did that either

but only normals started a phrase, broke it off to insert a prior event or a
comment on their word choice and then resumed the phrase as in:

(19) comment/correction (a) and then she-her father came home from work,
whatever-she asked her father for money.
(b) and a white-it appeared to be white-long sleeved
shirt
(c) so when her father came home-or the man who
came in the door I thought it was her father-came
in the door

2Professor Lambe and a lab assistant, Mary Paine.

3Chaika, Lambe, and Alexander's (in preparation) "Dysfluencies in normal and psychotic narra-
tions." will present a full statistical analysis of glitching. Here we are concerned with mutually
exclusive or virtually mutually exclusive error. Such disparity makes statistical analysis unnecessary.
322 CHAIKA AND ALEXANDER

In contrast, only psychotics started a phrase, broke off for a comment, but
never picked it up again and completed it. Examples of such psychotic omissions
are:

(20) syntactic gaps (a) he was blamed for and I didn't think that was fair the way the
way they did that either
(b) what are the and uh there was a scene
(c) and asks if she can have then goes to the ice cream place.
(d) but that's what kids are like these days have but th-it means
that [shinchuher] her parents that she's so proud of
[shinchuher] she goes out ... but she still is her destination is
unknown.

It must be emphasized that errors like those in (20) were not only exclusive to
psychotics, but the sentences in which they occurred were said as if nothing had
been omitted. There was no break in intonation or stress, but a vital word,
usually the head of a constituent structure, was never uttered. This constitutes a
true break in syntactic ability, a genuine agrammatism.
Although both populations evinced false starts, only psychotic false starts
seemed unrelated to the ultimate choice of word, as in

(21) false start (a) he ch- told where to go

in contrast to normal false starts which seemed to involve selecting the wrong
word in a set, erroneously choosing one pronoun over another, as in (21 b) and
starting to mention one kind of ice cream, a sundae, rather than the correct cone.

(b) she-we saw

(c) it looked like a chocolate su-a chocolate ice cream cone.

Only psychotics produced utterances in which a word which rhymed with an

apparent intended word was selected. There was no apparent recognition of these
slips. Examples are:

(22) rhyming slips (a) he twitched through the door

(b) that's all I can stew.

The twitch was probably intended to be switched as it was a reference to the

camera action, and the stew seems meant to be do as it ended the narrative.
There were very few syntactic errors, but both populations produced them.

(23) syntactic error (a) he charged it for her. J

(b) it's too close for dinnertime.
 PSYCHOTIC NARRATION 323

(c) . . . two three minutes for get waited on

(d) there was and when she got home
there was too near suppertime.

(23a), from a normal, is a simple reversal, and (b) and (c) are substitutions of for
for to. This is not as surprising as it might appear at first blush, as "for to"
1· together constitute the infinitive, as in, '' I would love for you to go.'' (23b) was
said by a normal and (23c) by a psychotic. The psychotic error (23d) also
1·
involves substitution of one word for a related one. Both there and it are dummy
subjects but, unlike for and to, they never occur in the same construction. All of
the errors in (23) were said with no apparent awareness that a slip had occurred.
Word salads, such a seemingly random throwing together of words that we
cannot discover their relationship to each other, not surprisingly occurred only in
psychotics. For example:

(24) word salad (a) hey if it we'd all in which is in not

Sometimes, word salads were difficult to separate out from neologisms in

which most of what appeared to be intended as a string of words are not words in
English. Because all of our psychotic subjects were monolingual, these could not
be attributed to foreign language interference.

(25) neologism (a) en ah ess pass by sh wu

(b) a little girl taking a dit
(c) she gets her ous ow after she . . .
(d) so therefore she etuh she ed she listened

Although neologizing is primarily a psychotic disruption, (25d) was produced by

a normal. This and other deviations in her narrative not surprisingly caused it to
be judged as psychotic by both raters.
Summing up this section, we see that only psychotics produce syntactic gaps,
rhyming slips, and word salads. With only one exception, they produced the
neologizing and the only false starts with elements unlike what followed. Only
normals produced self-corrections and comments on word choice, with no syn-
tactic or other gapping resulting. This is not to say that normals always picked up
on their own errors. Like psychotics, they often did not, but this occurred only
with single word errors in lexical items that may appear in the same construction.
When normals did pick up an error, they were still able to resume their narrative
without being deflected from the action of the videotape. Finally, normal errors,
for the most part, were not as damaging to comprehension as psychotic error
was. That is, the intended word was retrievable in normal narrative but not
always so in psychotic.
324 CHAIKA AND ALEXANDER

Chaika (1974) suggested that speech identified as schizophrenic arises from

an intermittent aphasia. That work has been roundly criticized by both psychol-
ogists and psychiatrists on the grounds that, as Cohen (1978, p. 1) said, "Hard
instances of agrammatism are hard to find" in such speech. It has been generally
assumed that, where schizophrenics (and manics) deviate in speech, that devia-
tion is on the discourse level. However, as this section shows, agrammatism does
occur in psychotic speech in the form of syntactic gaps (failure to provide the
head of a construction) and word salad. Moreover, such instances were not at all
difficult to find even in a population comprised solely of short-term patients.
All in all, these data support the conclusion that psychotic narratives are likely
to be less ordered at every level of narration than are normal ones. The next
section shows the disorganization in psychotic narratives that results in differing
patterns of misperceptions between the two populations.

MISPERCEPTIONS
It must be stressed that the very act of misperception does not define psychotics.
Actually, normals did about as much misperceiving as psychotics. The dif-
ferences lay both in the order of scanning the memory which seems to underlie
narrative production, and the kinds of misperceptions which each group had. As
will now be shown, this last was partially a result of the first.
The misperceptions of the two populations were almost mutually exclusive.
Misperceptions by normals arose out of their summing up the action so as to get a
smooth, logical progression of activity, all subordinated to what was apparently
seen as the central theme: the girl's desiring and then getting ice cream. Hence,
many normals, but not psychotics, reported that the father as well as the mother
refused the child. The story could be told either way. Normals ignored the
mother's affectionate and kind refusal. Rather, it became converted to a flat,
even unpleasant, denial of the girl's request. It was this scene that caused one
normal to say that the child was "rejected by one [parent]." Another reported
the mother as giving an abrupt "Nope." What is essential for the overall story
line is that the mother refuse, or else there would have been no reason for the girl
to ask her father. So, refuse she did-to the point of normals' grossly misprepre-
senting the character of the refusal. Similarly, it makes no difference to the story
exactly what flavor ice cream the girl gets, so normals did not seem to process
the clearly enunciated "double grape ice" in the videostory. As far as the central
story line goes, it makes no difference if the mother is preparing dinner rather
than setting the table. Consequently, a normal misperceived this sequence as
well.
Two normals misperceived the white cases barely visible through the window
which the child was looking into, so that one termed it a deli, and the other, a
-
I
PSYCHOTIC NARRATION 323

(c) . . . two three minutes for get waited on

(d) there was and when she got home
there was too near suppertime.

(23a), from a normal, is a simple reversal, and (b) and (c) are substitutions of for
for to. This is not as surprising as it might appear at first blush, as "for to"
together constitute the infinitive, as in, '' I would love for you to go.'' (23b) was
said by a normal and (23c) by a psychotic. The psychotic error (23d) also
involves substitution of one word for a related one. Both there and it are dummy
subjects but, unlike for and to, they never occur in the same construction. All of
the errors in (23) were said with no apparent awareness that a slip had occurred.
Word salads, such a seemingly random throwing together of words that we
cannot discover their relationship to each other, not surprisingly occurred only in
psychotics. For example:

(24) word salad (a) hey if it we'd all in which is in not

Sometimes, word salads were difficult to separate out from neologisms in

which most of what appeared to be intended as a string of words are not words in
English. Because all of our psychotic subjects were monolingual, these could not
be attributed to foreign language interference.

(25) neologism (a) en ah ess pass by sh wu

(b) a little girl taking a dit
(c) she gets her ous ow after she . . .
(d) so therefore she etuh she ed she listened

Although neologizing is primarily a psychotic disrJption, (25d) was produced by

a normal. This and other deviations in her narrative not surprisingly caused it to
be judged as psychotic by both raters.
Summing up this section, we see that only psychotics produce syntactic gaps,
rhyming slips, and word salads. With only one exception, they produced the
neologizing and the only false starts with elements unlike what followed. Only
normals produced self-corrections and comments on word choice, with no syn-
tactic or other gapping resulting. This is not to say that normals always picked up
on their own errors. Like psychotics, they often did not, but this occurred only
with single word errors in lexical items that may appear in the same construction.
When normals did pick up an error, they were still able to resume their narrative
without being deflected from the action of the videotape. Finally, normal errors,
for the most part, were not as damaging to comprehension as psychotic error
was. That is, the intended word was retrievable in normal narrative but not
always so in psychotic.
324 CHAIKA AND ALEXANDER

Chaika (1974) suggested that speech identified as schizophrenic arises from

an intermittent aphasia. That work has been roundly criticized by both psychol-
ogists and psychiatrists on the grounds that, as Cohen (1978, p. 1) said, "Hard
instances of agrammatism are hard to find" in such speech. It has been generally
assumed that, where schizophrenics (and manics) deviate in speech, that devia-
tion is on the discourse level. However, as this section shows, agrammatism does
occur in psychotic speech in the form of syntactic gaps (failure to provide the
head of a construction) and word salad. Moreover, such instances were not at all
difficult to find even in a population comprised solely of short-term patients.
All in all, these data support the conclusion that psychotic narratives are likely
to be less ordered at every level of narration than are normal ones. The next
section shows the disorganization in psychotic narratives that results in differing
patterns of misperceptions between the two populations.

MISPERCEPTIONS
It must be stressed that the very act of misperception does not define psychotics.
Actually, normals did about as much misperceiving as psychotics. The dif-
ferences lay both in the order of scanning the memory which seems to underlie
narrative production, and the kinds of misperceptions which each group had. As
will now be shown, this last was partially a result of the first.
The misperceptions of the two populations were almost mutually exclusive.
Misperceptions by normals arose out of their summing up the action so as to get a
smooth, logical progression of activity, all subordinated to what was apparently
seen as the central theme: the girl's desiring and then getting ice cream. Hence,
many normals, but not psychotics, reported that the father as well as the mother
refused the child. The story could be told either way. Normals ignored the
mother's affectionate and kind refusal. Rather, it became converted to a flat,
even unpleasant, denial of the girl's request. It was this scene that caused one
normal to say that the child was "rejected by one [parent]." Another reported
the mother as giving an abrupt •'Nope.'' What is essential for the overall story
line is that the mother refuse, or else there would have been no reason for the girl
to ask her father. So, refuse she did-to the point of normals' grossly misprepre-
senting the character of the refusal. Similarly, it makes no difference to the story
exactly what flavor ice cream the girl gets, so normals did not seem to process
the clearly enunciated "double grape ice" in the videostory. As far as the central
story line goes, it makes no difference if the mother is preparing dinner rather
than setting the table. Consequently, a normal misperceived this sequence as
well.
Two normals misperceived the white cases barely visible through the window
which the child was looking into, so that one termed it a deli, and the other, a
 PSYCHOTIC NARRATION 325

laundromat. Although the cases looked like either, it was surprising given the
entire videostory that they did not perceive that she was looking at ice cream
r cases, especially given the normals' penchant for fitting the facts to the perceived
story. However, even these errors did not mar the story line.
r In contrast to the clear evidence of normal misperception, psychotic miscod-
r ing of what they have seen is ambiguous. Given their difficulties in word finding,
it is not clear whether or not psychotics are producing lexical error or if they have

I misperceived. Certainly, we cannot find an overriding principle for error as we

can for normals. What seem to be psychotic misperceptions might also arise from
inappropriate focusing on associations not germane to the videostory.
Because we cannot ask these subjects what they meant,4 some apparently
visual errors, if not all, might be instances of word-retrieval error. For instance,
one manic narrator says the girl is looking in a trashcan. She is actually looking
into a window. It appears improbable that he was hallucinating about the
trashcan because he says, " ... This little girl was looking in the store was
looking in a trashcan or something." The' "or something" indicates that he was
not encoding a hallucinatory trashcan, but either that he could not recall what she
was actually looking into, or he had not registered that information. Alter-
natively, he might not have been able to think of the actual word window, so used
another noun which could complete "looking into." His "or something," like
"whatever," is frequently used to indicate that a word just selected is not quite
on target. Normal narrators have this problem as well, although they are not
likely to be as far off as trashcan for "window."
Again, when a psychotic encodes the child's request to her parents as, "She
talked to her mother and father," we do not know if he actually "saw" the child
conversing with, but not requesting anything from, her parents. Whereas re-
questing is a form of talking, still talking is not a usual synonym for it. It is as if
the subject hit upon a hyponym under which requesting or asking is categorized,
but did not quite get to his goal. Similarly, some psychotics spoke of the ''candy
store" rather than the ice cream store. Interesting! y, in Rhode Island, one does
not buy ice cream at a candy store. One purchases it variously at a creamery,
dairy, ice cream parlor, supermarket, variety store, or spa. Only candy, and
quality candy at that, is bought at a candy store. Every one of the subjects who
substituted candy store for ice cream parlor came from Rhode Island, as did all
but one who said the girl was going for a candy bar instead of an ice cream cone
(and that one might have been speaking of a hallucinatory girl with a baby). Both
of these, then, might be misperceptions, or they might be word-choice errors.

4 Psychotics are rarely able to explain their speech production, especially while in the throes of a

psychotic bout. Pointing out error to them, or asking them why they said something, does not usually
gain the questioner an appropriate answer.
326 CHAIKA AND ALEXANDER

CONCLUSION

Although it is generally conceded that many, although not necessarily all, psy-
chotics speak strangely, the etiology of this strangeness has long been in dispute.
One consistent problem with interpreting schizophrenic and manic speech has
been that, given its frequent opacity, and the fact that we may not know what it is
the psychotic is trying to say, we can not always tell if such speech is genuinely
creative, deliberately misleading, or involuntarily erroneous.
Providing subjects with a videostory about which they must form a narration
allows us to match speech production with its target. Even the most deviant
psychotic narratives in this study showed compliance with the task. Because
some psychotics do not evince structurally deviant speech, participants in this
study were preselected for invitation only if they did so.
It was found that psychotics produced error at almost every level of speech:
word formation, sentence production, and narrative production. Their speech
was characterized by a general deficit in ability to order and to organize. They
also lacked details of the videostory and used emotionally laden words. Patients
differed in the levels affected as well as in the severity of disruption, but the
general pattern was the same. Those normal narratives judged psychotic shared
one or more of these features, although no normal produced syntactic gaps, word
salad, rhyming slips of the tongue, or false starts with a sound different from the
target word. Only one normal produced neologisms. Nor did any normal produce
narratives on the basis of internal stimuli as did some psychotics. The psychotic
errors, then, are different both in degree and in kind from normal ones. This is an
important finding given the general consensus in medical and psychological
circles that such speech does not arise from a linguistic disability.
Normals organize their narratives far more tightly than do psychotics, fre-
quently utilizing temporal order and attempting to reproduce the details of what
they have seen. Although they do have minor linguistic errors, with one excep-
tion the target is easily retrieved by hearers. Not surprisingly, normals are both
capable of self-correction and likely to indulge in it. Where they err in reporting
events, they do so because they produce a coherent whole, so that they fit the
facts to what they perceive to be the central issues in the story.
Psychotics are not capable of such organization, therefore their mispercep-
tions do not form a coherent narrative. Their errors provide little evidence that
schizophrenics (or manics) produce "plus deviations" (Lecours & Vaniers-
Clement, 1976; Forrest, l 976.) Rather, almost all of their errors were "minus
deviations," hindering comprehension and/ or failing to encode the story. Be-
cause, in a variety of ways, they showed that they were trying to narrate the
story, one can conclude that they were not always able to say what they meant.
This argues for disrupted speaking skills.
Chafe's (1980) and Chaika's (1982a) hypotheses that narrative production
326 CHAIKA AND ALEXANDER

CONCLUSION

Although it is generally conceded that many, although not necessarily all, psy-
chotics speak strangely, the etiology of this strangeness has long been in dispute.
One consistent problem with interpreting schizophrenic and manic speech has
been that, given its frequent opacity, and the fact that we may not know what it is
the psychotic is trying to say, we can not always tell if such speech is genuinely
creative, deliberately misleading, or involuntarily erroneous.
Providing subjects with a videostory about which they must form a narration
allows us to match speech production with its target. Even the most deviant
psychotic narratives in this study showed compliance with the task. Because
some psychotics do not evince structurally deviant speech, participants in this
study were preselected for invitation only if they did so.
It was found that psychotics produced error at almost every level of speech:
word formation, sentence production, and narrative production. Their speech
was characterized by a general deficit in ability to order and to organize. They
also lacked details of the videostory and used emotionally laden words. Patients
differed in the levels affected as well as in the severity of disruption, but the
general pattern was the same. Those normal narratives judged psychotic shared
one or more of these features, although no normal produced syntactic gaps, word
salad, rhyming slips of the tongue, or false starts with a sound different from the
target word. Only one normal produced neologisms. Nor did any normal produce
narratives on the basis of internal stimuli as did some psychotics. The psychotic
errors, then, are different both in degree and in kind from normal ones. This is an
important finding given the general consensus in medical and psychological
circles that such speech does not arise from a linguistic disability.
Normals organize their narratives far more tightly than do psychotics, fre-
quently utilizing temporal order and attempting to reproduce the details of what
they have seen. Although they do have minor linguistic errors, with one excep-
tion the target is easily retrieved by hearers. Not surprisingly, normals are both
capable of self-correction and likely to indulge in it. Where they err in reporting
events, they do so because they produce a coherent whole, so that they fit the
facts to what they perceive to be the central issues in the story.
Psychotics are not capable of such organization, therefore their mispercep-
tions do not form a coherent narrative. Their errors provide little evidence that
schizophrenics (or manics) produce "plus deviations" (Lecours & Vaniers- ..,
Clement, 1976; Forrest, 1976.) Rather, almost all of their errors were "minus
deviations," hindering comprehension and/ or failing to encode the story. Be-
cause, in a variety of ways, they showed that they were trying to narrate the
story, one can conclude that they were not always able to say what they meant.
This argues for disrupted speaking skills.
Chafe's (1980) and Chaika's (1982a) hypotheses that narrative production
---
I
326 CHAIKA AND ALEXANDER

CONCLUSION

Although it is generally conceded that many, although not necessarily all, psy-
chotics speak strangely, the etiology of this strangeness has long been in dispute.
One consistent problem with interpreting schizophrenic and manic speech has
been that, given its frequent opacity, and the fact that we may not know what it is
the psychotic is trying to say, we can not always tell if such speech is genuinely
creative, deliberately misleading, or involuntarily erroneous.
Providing subjects with a videostory about which they must form a narration
allows us to match speech production with its target. Even the most deviant
psychotic narratives in this study showed compliance with the task. Because
some psychotics do not evince structurally deviant speech, participants in this
study were preselected for invitation only if they did so.
It was found that psychotics produced error at almost every level of speech:
word formation, sentence production, and narrative production. Their speech
was characterized by a general deficit in ability to order and to organize. They
also lacked details of the videostory and used emotionally laden words. Patients
differed in the levels affected as well as in the severity of disruption, but the
general pattern was the same. Those normal narratives judged psychotic shared
one or more of these features, although no normal produced syntactic gaps, word
salad, rhyming slips of the tongue, or false starts with a sound different from the
target word. Only one normal produced neologisms. Nor did any normal produce
narratives on the basis of internal stimuli as did some psychotics. The psychotic
errors, then, are different both in degree and in kind from normal ones. This is an
important finding given the general consensus in medical and psychological
circles that such speech does not arise from a linguistic disability.
Normals organize their narratives far more tightly than do psychotics, fre-
quently utilizing temporal order and attempting to reproduce the details of what
they have seen. Although they do have minor linguistic errors, with one excep-
tion the target is easily retrieved by hearers. Not surprisingly, normals are both
capable of self-correction and likely to indulge in it. Where they err in reporting
events, they do so because they produce a coherent whole, so that they fit the
facts to what they perceive to be the central issues in the story.
Psychotics are not capable of such organization, therefore their mispercep-
tions do not form a coherent narrative. Their errors provide little evidence that
schizophrenics (or manics) produce "plus deviations" (Lecours & Vaniers- ..,
Clement, 1976; Forrest, 1976.) Rather, almost all of their errors were "minus
deviations," hindering comprehension and/or failing to encode the story. Be-
cause, in a variety of ways, they showed that they were trying to narrate the
story, one can conclude that they were not always able to say what they meant.
This argues for disrupted speaking skills.
Chafe's (1980) and Chaika's (1982a) hypotheses that narrative production
 PSYCHITTIC NARRATION 327

parallels eye tracking seem to be supported by these data. So, too, are the works
of Oltmanns (1978), Neale and Oltmanns (1980), McGhie and Chapman (1969),
Reed (1970), and others who posit schizophrenic vulnerability to distraction and
faulty filtering mechanisms. The same phenomena have also been termed atten-
tional deficits.

REFERENCES
Alexander, P., VanKammer, D., & Bunney, W. (1979). Antipsychotic effects of lithium in schizo-
phrenia. Amercian Journal of Psychiatry, 136, 283-287.
Andreasen, N., & Powers, P. (1974). Overinclusive thinking in mania and schizophrenia. British
Journal of Psychiatry, 125, 452-456.
Bannister, D. (1962). The nature and measurement of schizophrenic thought disorder. Journal of
Mental Science, 108, 825-842.
Bernstein, B. (1971). Class, codes, and control (Vol. 1). London: Routledge & Kegan-Paul.
Broadbent, D.E. (1971). Decision and stress. New York: Academic Press.
Brown, R. (1973). Schizophrenia, language, and reality. American Psychologist, 28, 395-403.
Chafe, W. (Ed.). (1980). The Pear stories: Cognitive, cultural, and linguistic aspects of narrative
production. Norwood, NJ: Ablex.
Chaika, E. (1974). A linguist looks at "schizophrenic" language. Brain and Language, 1, 257-276.
Chaika, E. (1977). Schizophrenic speech, slips of the tongue, and jargonaphasis. Brain and Lan-
guage, 4, 464-475.
Chaika, E. (1981). How shall a discourse be understood? Discourse Processes, 4, 71-88.
Chaika, E. (1982a). A unified explanation for the diverse structural deviations reported for adult
schizophrenics with disrupted speech. Journal of Communication Disorders, 15, 167-189.
Chaika, E. (1982b). Thought disorder or speech disorder in schizophrenia? Schizophrenia Bulletin,
8, 588-591.
Chaika, E. (1982c). Language: The social mirror. Rowley, MA: Newbury House.
Chaika, E. (1983, December). Cohesive devices in normal and psychotic narrative. Paper delivered
at the annual meeting of the Linguistic Society of America, Minneapolis, MN.
Chaika, E., Lambe, R., & Alexander, P. (1986). Cohesive devices in normal and psychotic dis-
course. Manuscript submitted for publication.
Chapman, L.J., Chapman, J.P., & Daut, R. (1976). Schizophrenic inability to disattend from strong
aspects of meaning. Journal of Abnormal Psychology, 85, 35-40.
Cohen, B. (1978). Referent communication disturbances in schizophrenia. In S. Schwartz (Ed.),
Language and cognition in schizophrenia. Hillsdale, NJ: Erlbaum.
DiSimone, F.G., Darley, F., & Aronson, A. (1977). Patterns of dysfunction in schizophrenic
patients on an aphasia test battery. Journal of Hearing and Speech Disorders, 42, 498-513.
Forrest, D. (1%5). Poesis and the language of schizophrenia. Psychiatry, 28, 1-18.
Forrest, D. (1976). Nonsense and sense in schizophrenic language. Schizophrenia Bulletin, 2, 286-
298.
Fromkin, V. (1975). A linguist looks at "A linguist looks at 'schizophrenic' language." Brain and
Language, 2, 498-503.
Goody, E. (1978). Questions and politeness. New York: Cambridge University Press.
Holzman, P. (1978). Cognitive impairment and cognitive stability: Towards a theory of thought
disorder. In G. Serban (Ed.), Cognitive defects in the development of mental illness. New
York: Brunner/Maze!.
Kertesz, A. (1982). Schizophasia is distinct but not aphasic. The Behavioral and Brain Sciences, 4,
600.
328 CHAIKA AND ALEXANDER

Laffal, J. (1965). Pathological and normal language. New York: Atherton.

Laffal, J. (1982). Language competence and schizophrenic language. The Behavioral and Brain
Sciences, 4, 604-605.
Lanin-Kettering, I., and Harrow, M. (1985). The thought behind the words: A view of schizophrenic
speech and thinking disorders. Schizophrenia Bulletin, 11, l- 7.
Lecours, A.R., and Vaniers-Clement, M. (1976). Schizophrenia and jargonaphasia: A comparative
description with comments on Chaika's and Fromkin's respective looks at "schizophrenic"
language. Brain and Language, 3, 516-565.
Maher, B., McKeon, K., & McLaughlin, B. Studies in psychotic language. (1966). In P. Stone, D.
Dunphy, M. Smith, & D. Ogilvie (Eds.), General inquirer. Cambridge, MA: MIT Press.
Maher, B. (1972). The language of schizophrenia: A review and an interpretation. British Journal of
Psychiatry, 12, 3-17.
Martin, J. (1982). What is language? The Behavioral and Brain Sciences, 4, 607-608.
McGhie, A., & Chapman, J. (1969). Orders of attention in early schizophrenia. In A. H. Buss &
E. H. Buss (Eds.), Theories of schizophrenia. New York: Atherton Press.
Neale, J.M., & Oltmanns, T.F. (1980). Schizophrenia. New York: Wiley.
Oltmanns, T.F. (1978). Selective attention in schizophrenic and manic psychoses: The effect of
distraction on information processing. Journal of Abnormal Psychology, 87, 212-225.
Philips, S. (1976). Some sources of cultural variability in the regulation of talk. Language in Society,
5, 81-95.
Reed, J.L. (1970). Schizophrenic thought disorder: A review and hypothesis. Comprehensive Psy-
chiatry, 11, 403-432.
Rochester, S., Martin, J., & Thurston, S. (1977). Thought process disorder in schizophrenia: The
listener's task. Brain and Language, 4, 95-114.
Rochester, S., & Martin, J. (1979). Crazy talk: A study of the discourse of schizophrenic speakers.
New York: Plenum.
Sacks, H. (1964-1972). Lecture notes (mimeograph).
Schwartz, S. (1978). Language and cognition in schizrophrenia: A review and synthesis. In S.
Schwartz (Ed.), Language and cognition in schizophrenia. New York: Wiley.
Schwartz, S. (1982). Is there a schizophrenic language? The Behavioral and Brain Sciences, 4, 57.
Searle, J.R. (1983). Intentionality: An essay in the philosophy of mind. New York: Cambridge
Univesity Press.
Searles, H. (1967). The schizophrenic's experience of his world. Psychiatry, 30, ll9-l3l.
Tannen, D. (1980). A comparative analysis of oral narrative strategies: Athenian Greek and Ameri-
can English. In W. Chafe (Ed.), The Pear stories: Cognitive, cultural, and linguistic aspects
of narrative production. Norwood, NJ: Ablex.
Werner, 0., Lewis-Matichek, G., Evans, M., & Litowitz, B. (1975). An ethnosciences view of
schizophrenic speech. In M. Sances and B. Blounts (Eds.), Sociocultural dimensions of
language use. New York: Academic Press.

View publication stats